Cursor Bets $60 Billion on Orchestration Layer as AI Models Become Commodity

Cursor SDK Launch Signals End of IDE-Centric Era

Cursor has released its Software Development Kit (SDK) in public beta, packaging years of internal orchestration work into a model-agnostic tool for developers. The move confirms the company's belief that the value in AI coding lies not in models but in the harness that controls them.

CEO Michael Truell has called this the “third era” of AI software development, where agents working in cloud VMs will dominate. “More than a third of our internal pull requests are now created by agents,” Truell said. “Within a year, we expect the vast majority of development work to look that way.”

Cursor’s SDK (available via npm install @cursor/sdk) lets developers build agents directly on its harness, with support for codebase indexing, MCP servers, subagents, and observability hooks. It runs locally or on Cursor Cloud against dedicated virtual machines.

Background: From IDE to Agent Orchestrator

Twelve months ago, Cursor had 2.5 times more autocomplete users than agent users. Today, agent users outnumber tab users two to one. The company has spent two months signaling its shift beyond being an IDE company.

Last week, Cursor announced a partnership with SpaceX to train its proprietary Composer models on xAI’s Colossus supercomputer. In February, Truell published an essay declaring the “third era” of AI software development. The message is clear: Cursor thinks the AI model is becoming a commodity, and the product that wins the next decade is the harness around it.

External confirmation came from Google, which told The New Stack this week: “We don’t care which coding tool developers use—Gemini, Claude Code, or Cursor.” This reinforces that models themselves are interchangeable, while the orchestration layer adds lasting value.

What This Means for Developers and the Industry

Cursor’s SDK puts it in direct competition with OpenAI’s Agents SDK and Anthropic’s Claude Agent SDK. The battle is no longer about which AI model is best—it’s about which platform can best orchestrate agents to work reliably over hours, returning logs, video recordings, and live previews.

For developers, the implication is that the IDE becomes a fallback. According to Truell, agent usage at Cursor has grown more than 15x in the last year. The harness—not the model—will determine productivity gains. Workers who learn to harness AI agents will define the next era of their industries.

Cursor has placed a $60 billion bet on that proposition. The SDK is the first step in letting the entire industry build on Cursor’s orchestration backbone, independent of any single AI model provider.